Machine Learning with AMD GPUs

Photo by Lubo Minar on Unsplash

You want to run PyTorch or TensorFlow on your PC, but you have an AMD GPU. No worries, this guide will get you up and running in no time!
Use Ubuntu LTS to get started as soon as possible.

First, install ROCm, the AMD equivalent of CUDA.

https://rocmdocs.amd.com/en/latest/Installation_Guide/Installation-Guide.html

Then, install your favorite machine learning library.
The easiest way is to use docker.

For Pytorch, follow this:

https://hub.docker.com/r/rocm/pytorch

For Tensorflow, follow this:

https://hub.docker.com/r/rocm/tensorflow

Now to start development, I recommend using Visual Studio Code (VSC).

Method 1:

First, start Jupyter in your container, copy the URI.

http://localhost:8080/?token=cec254a8a1us02d44e3d466aa72bde7a303cd3b2463ef7b4

Now, configure your VSC to use the remote Jupyter server: for more details, follow this link:
https://code.visualstudio.com/docs/datascience/jupyter-notebooks#_connect-to-a-remote-jupyter-server

Method 2:

Or, install the “Remote — Containers” extension for VSC.

ext install ms-vscode-remote.remote-containers

Happy hacking!